Look2You

Inspiration

Online shopping gives people endless fashion inspiration, but it does not always help them understand what actually works for them. I often see an outfit on a model, influencer, or celebrity and want to recreate the entire look. However, after buying similar items, I realize that the silhouette, fit, or proportions do not suit my body. Many of those purchases end up sitting unworn in my closet.

This made me realize that people do not necessarily want an exact copy of an influencer’s outfit. What they really want is the same feeling—the color palette, mood, styling idea, or overall aesthetic—translated into something that works for their own body, lifestyle, budget, and preferences.

That insight inspired Look2You, an AI-powered personal stylist that transforms fashion inspiration into personalized, practical, and shoppable outfits.

What it does

Look2You helps users turn an outfit they like into a version that is better suited to them.

Users can provide information such as:

• Body proportions and fit concerns
• Preferred styles and colors
• Budget
• Occasion or daily activity
• Clothing preferences
• An inspiration image or outfit reference

Gemini analyzes both the inspiration look and the user’s personal context. Instead of simply recommending identical products, Look2You identifies the elements that make the original outfit appealing, such as its color palette, silhouette, layering, mood, and level of formality.

It then adapts those elements to the individual user. For example, if a user with broad shoulders likes an outfit with a structured or shoulder-emphasizing top, Look2You can preserve the original color palette and overall aesthetic while recommending a more flattering neckline, fabric, or silhouette.

The system generates several personalized outfit options, explains why each look may work for the user, and provides relevant shopping recommendations.

How we built it

We designed Look2You around a multimodal AI workflow.

First, the user submits an inspiration image along with personal information about their body, preferences, budget, and intended occasion.

Gemini then analyzes the image to identify key fashion attributes, including:

• Garment categories
• Colors and patterns
• Silhouettes and proportions
• Materials and textures
• Styling techniques
• Overall aesthetic and occasion

Next, the system combines the visual analysis with the user profile. It determines which elements of the original outfit should be preserved and which elements should be adjusted.

Finally, Look2You generates personalized outfit recommendations with explanations and suggested products. The goal is not to judge or score the user’s body, but to provide supportive, practical guidance that helps the user make more confident purchasing decisions.

Challenges we faced

One of the biggest challenges was translating subjective fashion preferences into structured information that an AI system could understand. A user may say that an outfit “feels elegant” or “looks effortless,” but those impressions can come from many different elements, such as color, proportion, texture, or styling.

Another challenge was balancing inspiration with personalization. Copying an outfit exactly is relatively simple, but preserving its visual identity while changing the fit, silhouette, or individual garments requires deeper reasoning.

We also had to think carefully about how to discuss body shape. We wanted the experience to feel encouraging rather than judgmental, so we focused on fit preferences, comfort, and personal goals instead of assigning negative labels or numerical attractiveness scores.

Product availability, sizing differences between brands, and keeping shopping recommendations within a user’s budget were also important challenges.

What we learned

We learned that personalization in fashion requires more than collecting height, weight, or clothing size. A useful recommendation must also consider lifestyle, confidence, comfort, occasion, budget, and the emotional reason a user was attracted to an outfit in the first place.

We also learned that multimodal AI is especially powerful for this problem because fashion inspiration is highly visual. Gemini can help connect what the user sees in an image with what they describe in words.

Most importantly, we learned that the best AI stylist should not tell users what they are allowed to wear. It should help them understand their preferences, explore alternatives, and make decisions with greater confidence.

What's next for Look2You

Our long-term goal is to help users gradually develop a personal style profile that improves over time.

Future features may include:

• Virtual try-on and outfit visualization
• Personalized size recommendations
• Real-time product inventory and availability
• Weather-aware and calendar-aware outfit planning
• Recommendations based on items already in the user’s closet
• Feedback loops that learn from saved, purchased, returned, and frequently worn items
• Social media inspiration analysis
• Weekly outfit planning for work, school, travel, and special occasions

Ultimately, Look2You aims to reduce impulsive purchases and help users build a wardrobe that feels personal, wearable, and genuinely useful.

Built With

• and
• fastapi
• firebase
• gemini-api
• google-ai-studio
• google-cloud
• python
• react
• tailwind-css
• typescript